An hommage to Alain Desrosières, statistician, sociologist, historian and philosopher of statistics

By Françoise Milewski and  Henri Sterdyniak

Alain Desrosières has passed away, at the age of 72. An administrator at the INSEE, he had been editor of the journal Économie et statistique, then head of the Department of social studies, before working on the comparative analysis of Europe’s statistical systems.

He was the troubled conscience of official statistics in France.

Alain’s many books and articles traced the birth and growth of statistics. His articles discuss their scientific and social foundations. They highlight the links between statistical standards and the production of statistics, between the history of economic policy and statistical methods and categories, in the face of the trend to “naturalize” them. “The ways of thinking society, managing it and quantifying it are inseparable”, he declared. Statistics cannot be separated from its use, and it evolves with changes in public policy. And so, for instance, he raised questions about “the quality of quantity”.

Alain passionately lived and studied the contradictions of statistics, a tool for knowledge and a tool for governing. Are statistics in the service of democracy, helping society to better understand itself, or of the State, helping it to better achieve its goals? And this State, which organizes and finances the statistical system, itself has two faces: the welfare state, an instrument of resistance to market forces, as well as a State in the service of a social formation shaped by capitalism.

Statistics measures and classifies. But is it a neutral scientific discipline, or does it express the vision that society has of itself at a given point, especially since it must rely on administrative sources that are themselves not neutral? Should it base itself on people’s everyday experience, or, on the contrary, challenge this in the name of science?

Can we account for different societies using the same categories? Alain has devoted great attention to the statistical harmonization that the European Union implies, with its risk of negating differences between societies.

He questioned the policy on indicators implemented by the Open Method of Coordination (OMC) and France’s organic law on budget bills (LOLF). Policies define indicators that statisticians are supposed to measure, and then set targets for these indicators. But this practice is dangerous, as these indicators become the focus of the analysis even as the policies aim to improve the indicators, which tends to cause them to lose their significance.

Below we reproduce some snippets from his articles, as an invitation to read them in their entirety. The myth of the data that is indisputable because impartial, the unconditional respect in the face of indicators that, because quantified are thus indisputable, regardless of the methods, standards and conventions underpinning their calculation – all these are a constant threat for the social sciences, particularly economics. And for society.

Alain Desrosières took part in numerous meetings of statisticians in order to give his colleagues food for thought about their practices and their methods (see in particular the conference of 30 March 2011: “Official statistics as a unique public good“, Workshop 3). He developed fertile links between statistical practice and sociologists, in particular Pierre Bourdieu and Bruno Latour.

He showed the influence of nomenclatures on the constitution of statistical information and, through that, on the structuring of society (Les Catégories socioprofessionnelles, co-authored by Laurent Thévenot, La Découverte, Repères collection, 1988).

Alain leaves us a number of major works: La politique des grands nombres, histoire de la raison statistique (Editions La Découverte, Paris, 1993) and L’argument statistique, in two volumes: I: Pour une sociologie historique de la quantification, and II: Gouverner par les nombres (Les Presses des Mines ParisTech, Sciences sociales collection, Paris, 2008).

He leaves us his most recent work: “Est-il bon, est-il méchant ? Le rôle du nombre dans le gouvernement de la cité néolibérale” (Nouvelles perspectives en sciences sociales, volume 7, no. 2, May 2012).

Alain set an example as a modest but demanding intellectual who sought to put his professional experience and scientific efforts in the service of democracy.


A few short excerpts from his writings:

“How can the contradiction be resolved between the ethos of the statisticians and taking feedback into account, even when it seems to them just an annoying obstacle to their mission, which they conceive of as ‘providing unbiased reflections of reality’? It is not possible to isolate a moment of measurement that is independent of its uses, in particular the conventions that are the first step in quantification. The training of statisticians needs to be decompartmentalized and supplemented with the study of history, political science, the sociology of statistics, econometrics, probability, accounting and management. This program, inspired by the achievements of Sciences Studies (Pestre, 2006), could facilitate the inclusion of quantitative tools in social debates, without winding up in either a priori rejection or unconditional, naïve respect for ‘facts that are indisputable because quantified’.”

Est-il bon, est-il méchant ? Le rôle du nombre dans la cité néolibérale. Conclusion of a presentation to the seminar L’Informazione Prima Dell’Informazione. Conoscenza E Scelte Pubbliche, Milan Bicocca, 27 May 2010, Nouvelles perspectives en sciences sociales, volume 7, no. 2, May 2012.


“Quantification has become a sign of objectivity, rigor and impartiality that is mobilized in a variety of situations, from political debate to scientific demonstration, and including business indicators and the measurement of public opinion. However, quantification, in its various statistical formats, is not content merely to provide a reflection of the world, but also creates new ways of thinking, representing, expressing and acting on it, through the power of its models and its procedures, its broad dissemination and its use in argumentation. This book shows how ‘statistical argument’ is historically constructed, and what the cognitive and social effects of quantification systems are today.”

Pour une sociologie historique de la quantification, Volume 1 of L’argument statistique (Les Presses des Mines Paris-tech, Sciences socials collection, Paris, 2008), back cover.


“Governments of men use and abuse the ‘argument of statistics’. With the emergence of a neo-liberal state, public policy is increasingly relying on quantitative indicators that provide evaluations of the performance of different policy actions. The various ‘winners’ are broadcast widely (often under the Anglo-American rubric of ‘benchmarking’), ranking high schools, universities, even nations. This rite of quantification, far from providing a neutral image of phenomena, transforms and performs them. This book offers specific case studies, surveys of family budgets, planning commissions, local statistics and national accounts, analyzing the production of official statistics and their use by the public authorities. And it will be seen how statistics has imposed itself as both an evidentiary tool in the empirical sciences and a tool of government, in accordance with the intuition that Foucault had already presented in the 1970s under the name of ‘governmentality’.”

Gouverner par les nombres, Volume 2 of L’argument statistique (Les Presses des Mines Paris-tech, Sciences socials collection, Paris, 2008), back cover.


“Major crises are of course times when statistics are mobilized intensively to express the gravity of the situation. But they are also times of great debate, during which the role of the state in the regulation and control of the economy is completely rethought. To each of these crises corresponds the emergence of new ways of quantifying the social world. New models of action imply new variables and new systems of observation.

Economic and political history from the 1880s to the present day has offered at least three (if not four) examples of such configurations, combining ways of thinking society, ways of acting on it, and statistics adapted to the times. The crisis of the 1880s prompted the great statistics on labour and employment. The crisis of 1929 was the source of Keynesian macroeconomic policies and national accounts. The crisis of the 1970s was thought about in the neoliberal categories of microeconomics, and led to state reforms focusing in particular on performance indicators. Finally, the two crises of the 2000s, ecological and then financial, will perhaps give rise to radically new ways of thinking and quantifying public action. A review of the way that a few somewhat older crises were experienced, and their impact on the use of official statistics, may be useful for thinking about the magnitude of the changes that may result from these two recent crises.”

Crises économiques et statistiques, de 1880 à 2010“, ParisTech Review, 30 August 2010.